Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A computer-implemented method comprising: receiving, from a device, input audio data corresponding to an utterance; performing speech processing on the input audio data to generate first intent data representing the utterance; determining first output data based at least in part on the first intent data, the first output data being responsive to the utterance; receiving context data corresponding to the utterance; based at least in part on the context data, determining additional content is to be output, wherein the additional content is nonresponsive to the utterance; generating second intent data representing the additional content; determining second output data based at least in part on the second intent data, the second output data corresponding to first additional content; causing the device to present first content corresponding to the first output data; and causing the device to present the first additional content.
This invention relates to a computer-implemented method for enhancing voice-based interactions by dynamically generating context-aware responses. The method addresses the limitation of traditional voice assistants, which typically provide only direct responses to user queries without leveraging contextual information to offer additional relevant content. The method begins by receiving input audio data from a device, such as a smartphone or smart speaker, corresponding to a user's spoken utterance. The system processes this audio data using speech recognition to generate intent data representing the user's request or command. Based on this intent data, the system determines an initial output response that directly addresses the user's utterance. Additionally, the system receives context data related to the utterance, which may include factors such as the user's location, time of day, recent interactions, or preferences. Using this context, the system identifies additional content that is relevant but not directly responsive to the user's query. For example, if a user asks for a weather forecast, the system might also provide nearby event recommendations based on the user's location and calendar. The system then generates intent data for this additional content and determines corresponding output data. Both the primary response and the additional content are then presented to the user via the device. This approach improves user experience by proactively offering useful information beyond the immediate query, making interactions more intuitive and personalized.
2. The computer-implemented method of claim 1 , further comprising: determining previous interaction data associated with the device; and determining the previous interaction data indicates a pattern whereby a second utterance, represented by the second intent data, is received within a time threshold following output of content responsive to the utterance, wherein determining the second output data is further based at least in part on the previous interaction data indicating the pattern.
This invention relates to a computer-implemented method for processing user interactions with a voice-based or conversational interface system. The method addresses the problem of improving the relevance and efficiency of responses in multi-turn conversations by leveraging historical interaction patterns. The method involves receiving an utterance from a user device, converting the utterance into intent data representing the user's intent, and generating output data in response. The method further includes analyzing previous interaction data associated with the device to identify patterns where a second utterance, corresponding to a second intent, is received within a specific time threshold after the system outputs content in response to the initial utterance. If such a pattern is detected, the system uses this historical data to influence the generation of the second output data, ensuring more contextually appropriate and timely responses. This approach enhances user experience by anticipating follow-up interactions and reducing redundant or disjointed responses. The method may also involve tracking user preferences, interaction frequency, and contextual cues to refine future responses dynamically. The system can be applied in virtual assistants, chatbots, or other conversational AI applications to improve natural and efficient dialogue flow.
3. The computer-implemented method of claim 1 , further comprising: generating multi-dimensional vector data based at least in part on the context data; determining the multi-dimensional vector data includes at least a threshold amount of information; and based at least in part on the multi-dimensional vector data including at least the threshold amount of information, running a machine learned model that determines when additional content is to be output.
This invention relates to a computer-implemented method for processing context data to determine when to output additional content using machine learning. The method addresses the challenge of efficiently analyzing context data to make informed decisions about content delivery, ensuring relevance and timeliness. The system first generates multi-dimensional vector data from the context data, which represents the information in a structured format suitable for machine learning analysis. The method then evaluates whether this vector data contains at least a threshold amount of meaningful information, ensuring the data is sufficiently robust for reliable processing. If the threshold is met, a machine-learned model is executed to analyze the vector data and determine whether additional content should be output. The model assesses the context and makes a decision based on learned patterns, optimizing content delivery for user engagement or system efficiency. This approach enhances decision-making by leveraging structured data representation and machine learning, improving the accuracy and relevance of content output. The method is particularly useful in applications requiring dynamic content delivery, such as personalized recommendations, automated responses, or adaptive interfaces.
4. The computer-implemented method of claim 1 , further comprising: receiving, from the device, second input audio data corresponding to a second utterance; performing speech processing on the second input audio data to generate second intent data representing the second utterance; determining third output data based at least in part on the second intent data, the third output data being responsive to the second utterance; causing the device to present second content corresponding to at least a portion of the third output data; receiving, from the device, third input audio data corresponding to a third utterance; performing speech processing on the third input audio data to generate third intent data representing a first frequency, at which additional content is to be output, is to be decreased; determining a second frequency at which additional content has been output; determining the second frequency is greater than the first frequency; and based at least in part on the second frequency being greater than the first frequency, determining second additional content, nonresponsive to the second utterance, is not to be output.
This invention relates to a computer-implemented method for managing audio interactions in a voice-based system, addressing the problem of excessive or irrelevant content output during user interactions. The method involves receiving input audio data from a device, such as a smart speaker or virtual assistant, corresponding to user utterances. Speech processing is performed on the input audio data to generate intent data representing the user's spoken commands or queries. Based on this intent data, output data is generated to produce a response, which is then presented to the user via the device. The method further includes receiving subsequent input audio data corresponding to additional user utterances. Speech processing is applied to this data to generate new intent data, which is used to determine responsive output data. The device then presents content based on this output. Additionally, the method monitors the frequency at which additional content is output. If a user requests a decrease in the output frequency, the system compares this requested frequency with the actual output frequency. If the actual frequency exceeds the requested frequency, the system suppresses nonresponsive additional content to reduce information overload and improve user experience. This adaptive control ensures that the system dynamically adjusts content delivery based on user preferences, enhancing efficiency and relevance.
5. A system comprising: at least one processor; and at least one memory including instructions that, when executed by the at least one processor, cause the system to: receive, from a device, input data corresponding to a first command; determine first intent data representing the first command; determine first output data based at least in part on the first intent data, the first output data being responsive to the first command; receive context data corresponding to the first command; based at least in part on the context data, determine additional content is to be output, wherein the additional content is nonresponsive to the first command; determine second output data corresponding to first additional content; cause the device to present first content corresponding to the first output data; and cause the device to present the first additional content.
This system operates in the domain of intelligent user interfaces, particularly for devices that process natural language commands and provide context-aware responses. The problem addressed is the limitation of traditional command-response systems, which only provide direct answers to user queries without leveraging contextual information to offer additional relevant content. The system includes at least one processor and memory storing instructions that, when executed, enable the system to receive input data from a device, such as a voice assistant or chatbot, corresponding to a user command. The system processes this input to determine the user's intent and generates a primary response based on that intent. Additionally, the system receives context data related to the command, such as user preferences, historical interactions, or environmental factors. Using this context, the system identifies and determines additional content that is not directly responsive to the command but is relevant to the user. For example, if a user asks for a weather forecast, the system might also provide related content like local event recommendations based on the forecast. The system then generates output data for both the primary response and the additional content, causing the device to present both to the user. This approach enhances user experience by providing proactive, contextually relevant information beyond the immediate query.
6. The system of claim 5 , wherein the at least one memory further includes instructions that, when executed by the at least one processor, further cause the system to: determine previous interaction data associated with the device; and determine the previous interaction data indicates a pattern whereby second input data, corresponding to the first additional content, has been received following output of content responsive to the first command, wherein determining the second output data is based at least in part on the previous interaction data including the pattern.
A system for personalized content delivery analyzes user interactions with a device to predict and provide relevant additional content. The system processes input data from a device, such as a command or query, and generates output data, such as a response or content. The system further includes memory storing instructions that, when executed, cause the system to determine previous interaction data associated with the device. This data is analyzed to identify patterns where specific input data, corresponding to additional content, has been received after the system outputs content in response to a prior command. The system then uses this pattern to determine the second output data, ensuring the additional content is contextually relevant based on the user's historical behavior. The system dynamically adapts its responses by leveraging past interactions to improve the accuracy and relevance of the content provided, enhancing user experience by anticipating and delivering desired information. This approach reduces the need for repetitive user inputs and streamlines interaction efficiency.
7. The system of claim 5 , wherein the at least one memory further includes instructions that, when executed by the at least one processor, further cause the system to: generate multi-dimensional vector data based at least in part on the context data; determine the multi-dimensional vector data includes at least a threshold amount of information; and based at least in part on the multi-dimensional vector data including at least the threshold amount of information, run a machine learned model that determines when additional content is to be output.
This invention relates to a system for processing context data to determine when to output additional content using machine learning. The system addresses the challenge of efficiently analyzing contextual information to make dynamic content delivery decisions. The system includes at least one processor and at least one memory storing instructions that, when executed, cause the system to generate multi-dimensional vector data from context data. The system then evaluates whether this vector data contains at least a threshold amount of meaningful information. If the threshold is met, the system activates a machine-learned model to assess whether additional content should be output. The machine-learned model processes the vector data to make this determination. The system may also include components for collecting context data, such as user interactions, environmental conditions, or device states, and preprocessing this data into a structured format suitable for vectorization. The machine-learned model is trained to recognize patterns in the vector data that correlate with the need for additional content, ensuring timely and relevant content delivery. This approach improves user engagement by dynamically adapting content output based on real-time contextual analysis.
8. The system of claim 5 , wherein the at least one memory further includes instructions that, when executed by the at least one processor, further cause the system to: identify profile data associated with the input data; and determine at least a portion of the profile data indicating a maximum number of times the first additional content is to be output over a period of time, wherein the device is caused to present the first additional content based at least in part on the at least a portion of the profile data.
A system for managing content presentation includes a processor and memory storing instructions that, when executed, cause the system to identify profile data associated with input data and determine a maximum number of times additional content can be output over a specified period. The system then presents the additional content based on this profile data. The profile data may include user preferences, historical interaction data, or other contextual information that influences content delivery. The system ensures that the additional content is displayed or played within predefined limits to avoid over-exposure, enhancing user experience and engagement. The instructions may also cause the system to analyze the input data to select the additional content, ensuring relevance and personalization. The system dynamically adjusts content presentation based on real-time data, optimizing delivery while respecting user-defined or system-imposed constraints. This approach improves content management by balancing user preferences with operational limits, ensuring a controlled and tailored presentation of additional content.
9. The system of claim 5 , wherein the at least one memory further includes instructions that, when executed by the at least one processor, further cause the system to: identify profile data associated with the input data; and determine at least a portion of the profile data indicating a first time period when additional content is permitted to be output, wherein the context data includes a first portion representing a current time, wherein determining the additional content is to be output is further based at least in part on the at least a portion of the profile data and the first portion.
A system for managing content output based on user profiles and contextual factors. The system processes input data, such as user requests or environmental conditions, and determines whether to output additional content, such as advertisements or notifications. The system includes a processor and memory storing instructions that, when executed, analyze context data, including the current time, to decide whether to output additional content. The system also identifies profile data associated with the input data, which may include user preferences, restrictions, or permissions. The profile data specifies time periods when additional content is allowed or restricted. The system compares the current time from the context data with the permitted time periods in the profile data to determine whether to output the additional content. This ensures content is delivered only during authorized time frames, enhancing user experience and compliance with restrictions. The system may also integrate with other components, such as data processing modules or user interface systems, to refine content delivery decisions based on additional contextual or user-specific factors.
10. The system of claim 5 , wherein the at least one memory further includes instructions that, when executed by the at least one processor, further cause the system to: determine first resolved entity data representing an entity indicated in the first command; identify profile data associated with the input data; and determine at least a portion of the profile data indicating additional content is permitted to be output when a command corresponds to the first resolved entity data, wherein determining the additional content is to be output is further based at least in part on: the first resolved entity data representing the entity indicated in the first command, and the at least a portion of the profile data.
This invention relates to a system for managing and outputting content based on user commands and profile data. The system processes input data, such as voice or text commands, to determine the intended entity or action. It then resolves the entity data from the command and retrieves associated profile data linked to the input data. The profile data includes permissions or rules that dictate whether additional content can be output when a command corresponds to a specific resolved entity. The system evaluates the resolved entity data and the profile data to decide if supplementary content should be displayed or provided. This ensures that content output is contextually relevant and adheres to predefined permissions, enhancing user experience and security. The system dynamically adjusts content delivery based on the resolved entity and user-specific profile settings, allowing for personalized and controlled interactions. This approach improves the accuracy and appropriateness of content responses in interactive systems.
11. The system of claim 5 , wherein the at least one memory further includes instructions that, when executed by the at least one processor, further cause the system to: receive the first output data from a first application; and receive the second output data from a second application.
A system for processing data from multiple applications includes at least one processor and at least one memory storing instructions. The system is configured to receive first output data from a first application and second output data from a second application. The system processes these outputs to generate a combined result, which may involve analyzing, transforming, or integrating the data from both applications. The system may also include a user interface for displaying the combined result or other processed data. The memory further stores instructions for additional functions, such as data validation, error handling, or communication with external systems. The system ensures seamless interaction between different applications, improving data consistency and reducing manual intervention. This approach enhances efficiency in environments where multiple applications generate related but separate data streams. The system may be used in industries like finance, healthcare, or manufacturing, where integrating data from disparate sources is critical for decision-making. The instructions in the memory enable the processor to execute these functions autonomously, minimizing the need for manual configuration. The system's modular design allows for scalability, accommodating additional applications or data sources as needed.
12. The system of claim 5 , wherein the at least one memory further includes instructions that, when executed by the at least one processor, further cause the system to: receive, from the device, second input data corresponding to a second command; determine second intent data representing the second command; determine third output data based at least in part on the second intent data, the third output data being responsive to the second command; cause the device to present second content corresponding to at least a portion of the third output data; receive, from the device, third input data corresponding to a third command; determine third intent data representing the third command, the third intent data representing a first frequency, at which additional content is to be output, is to be decreased; determine a second frequency at which additional content has been output; determine the second frequency is greater than the first frequency; and based at least in part on the second frequency being greater than the first frequency, determine second additional content, nonresponsive to the third command, is not to be output.
This invention relates to a system for processing user commands and generating responsive content, particularly in scenarios where the system must adapt its behavior based on user interactions. The system includes at least one processor and memory storing instructions that, when executed, enable the system to receive input data from a device, such as a user command, and determine intent data representing the command. The system then generates output data responsive to the command and causes the device to present corresponding content. The system further processes subsequent commands, determining intent data for each and generating responsive output. Notably, the system can adjust its behavior based on user preferences or interaction patterns. For example, if a user issues a command to decrease the frequency of additional content (e.g., notifications or suggestions), the system evaluates whether the current output frequency exceeds the desired frequency. If so, the system suppresses non-responsive additional content to comply with the user's preference. This adaptive behavior ensures the system remains responsive to user needs while minimizing unnecessary or disruptive outputs. The system may be used in applications such as virtual assistants, smart home devices, or other interactive computing environments where dynamic user preferences must be respected.
13. A computer-implemented method comprising: receiving, from a device, input data corresponding to a first command; determining first intent data representing the first command; determining first output data based at least in part on the first intent data, the first output data being responsive to the first command; receiving context data corresponding to the first command; based at least in part on the context data, determining additional content is to be output, wherein the additional content is nonresponsive to the first command; determining second output data corresponding to first additional content; causing the device to present first content corresponding to the first output data; and causing the device to present the first additional content.
This invention relates to a computer-implemented method for processing user commands in a conversational interface, such as a virtual assistant or chatbot, to provide both responsive and contextually relevant additional content. The method addresses the problem of delivering a more engaging and informative user experience by dynamically generating supplementary information beyond the direct response to a user's query. The method involves receiving input data from a device, such as a voice command or text input, and analyzing it to determine the user's intent. Based on this intent, the system generates primary output data that directly answers or fulfills the user's command. Additionally, the system receives context data associated with the command, which may include user preferences, historical interactions, or environmental factors. Using this context, the system determines whether to provide additional content that is not directly responsive to the command but is relevant to the user's broader needs or interests. The system then generates secondary output data corresponding to this additional content and presents both the primary response and the supplementary content to the user. This approach enhances user engagement by offering proactive, context-aware suggestions or information.
14. The computer-implemented method of claim 13 , further comprising determining previous interaction data associated with the device; and determining the previous interaction data indicates a pattern whereby second input data, corresponding to the first additional content, has been received following output of content responsive to the first command, wherein determining the second output data is based at least in part on the previous interaction data including the pattern.
This invention relates to a computer-implemented method for enhancing user interactions with a device by leveraging historical interaction patterns. The method addresses the problem of inefficient or ineffective content delivery by analyzing past user behavior to predict and provide more relevant additional content. The method involves receiving a first command from a user, which triggers the output of content responsive to that command. The system then determines previous interaction data associated with the device, specifically looking for patterns where second input data—corresponding to additional content—was received after the initial content output. If such a pattern is identified, the system uses this historical data to determine second output data, which is additional content tailored to the user's likely needs based on their past behavior. This ensures that the device proactively provides relevant information without requiring explicit user requests, improving efficiency and user experience. The method may also involve generating a second command based on the second output data, which can further refine the content delivered to the user. By analyzing and applying these interaction patterns, the system dynamically adapts to user preferences, reducing redundancy and enhancing responsiveness. This approach is particularly useful in applications where user behavior is repetitive or predictable, such as smart assistants, recommendation systems, or automated customer service platforms.
15. The computer-implemented method of claim 13 , further comprising: generating multi-dimensional vector data based at least in part on the context data; determining the multi-dimensional vector data includes at least a threshold amount of information; and based at least in part on the multi-dimensional vector data including at least the threshold amount of information, running a machine learned model that determines when additional content is to be output.
The invention relates to a computer-implemented system for processing context data to dynamically control content output using machine learning. The method involves generating multi-dimensional vector representations from input context data, which may include user interactions, environmental factors, or other relevant signals. These vectors encode the contextual information into a structured format suitable for computational analysis. The system evaluates whether the generated vector data contains a sufficient amount of meaningful information by comparing it against a predefined threshold. This threshold ensures that only sufficiently informative data is used for downstream processing, preventing the model from making decisions based on sparse or irrelevant inputs. Once the vector data meets or exceeds the threshold, a trained machine learning model is executed to determine whether additional content should be presented to the user. The machine learning model analyzes the vectorized context to predict optimal content delivery timing, relevance, or necessity, thereby enhancing user engagement or system efficiency. The approach leverages vector-based data representation to capture complex relationships within the context, enabling more accurate and adaptive decision-making compared to traditional rule-based systems. The method is particularly suited for applications in recommendation systems, adaptive user interfaces, or real-time content delivery platforms where contextual awareness is critical.
16. The computer-implemented method of claim 13 , further comprising: identifying profile data associated with the input data; and determining at least a portion of the profile data indicating a maximum number of times the first additional content is to be output over a period of time, wherein the device is caused to present the first additional content based at least in part on the at least a portion of the profile data.
This invention relates to a computer-implemented method for controlling the presentation of additional content, such as advertisements, based on user profile data. The method addresses the problem of over-exposure to repetitive content, which can lead to user fatigue and reduced engagement. The system identifies profile data associated with input data, such as user behavior or preferences, and determines a maximum number of times additional content (e.g., an advertisement) should be displayed to a user over a specified period. The device then presents the content based on this limit, ensuring that the user is not overwhelmed by repeated exposure. The method may also involve tracking user interactions with the content to refine future presentations. By dynamically adjusting content delivery based on user-specific constraints, the system improves user experience and content effectiveness. The invention is particularly useful in digital advertising, personalized recommendations, and content management systems where controlled exposure is critical.
17. The computer-implemented method of claim 13 , further comprising: identifying profile data associated with the input data; and determining at least a portion of the profile data indicating a first time period when additional content is permitted to be output, wherein the context data includes a first portion representing a current time, wherein determining the additional content is to be output is further based at least in part on the at least a portion of the profile data and the first portion.
This invention relates to a computer-implemented method for controlling the output of additional content based on user profile data and contextual timing information. The method addresses the problem of determining when to display or output supplementary content, such as advertisements, notifications, or other media, in a way that respects user preferences and contextual constraints. The method involves analyzing input data, which may include user interactions, device states, or environmental conditions, to generate context data. This context data includes a timestamp representing the current time. The method further identifies profile data associated with the input data, which may include user preferences, settings, or historical behavior patterns. The profile data specifies at least one time period during which additional content is permitted to be output. The decision to output the additional content is based on both the context data (including the current time) and the relevant portion of the profile data. This ensures that the content is only displayed during authorized time windows, enhancing user experience and compliance with user preferences. The method may also involve filtering or prioritizing content based on additional criteria, such as user activity or device capabilities.
18. The computer-implemented method of claim 13 , further comprising: determining first resolved entity data representing an entity indicated in the first command; identifying profile data associated with the input data; and determining at least a portion of the profile data indicating additional content is permitted to be output when a command corresponds to the first resolved entity data, wherein determining the additional content is to be output is further based at least in part on: the first resolved entity data representing the entity indicated in the first command, and the at least a portion of the profile data.
This invention relates to a computer-implemented method for processing commands in a digital assistant or similar system, focusing on dynamically determining whether additional content should be output based on user profile data and resolved entity data. The method addresses the challenge of providing relevant and contextually appropriate responses by evaluating both the intent of a user's command and their profile preferences. The method involves receiving a first command from a user and determining first resolved entity data representing an entity indicated in the command. For example, if the command is "play music by Artist X," the resolved entity data would identify "Artist X." The system then identifies profile data associated with the user's input, which may include preferences, permissions, or historical behavior. The method checks whether the profile data permits additional content to be output when a command corresponds to the resolved entity data. For instance, if the user's profile allows recommendations, the system may suggest related artists or songs. The decision to output additional content is based on both the resolved entity data and the relevant portion of the profile data, ensuring the response aligns with the user's preferences and the context of the command. This approach enhances user experience by delivering personalized and contextually relevant information.
19. The computer-implemented method of claim 13 , further comprising: receiving the first output data from a first application; and receiving the second output data from a second application.
This invention relates to a computer-implemented method for processing output data from multiple applications. The method addresses the challenge of integrating and managing data streams from different software applications, which often operate independently and generate output in incompatible formats. The invention provides a solution by receiving first output data from a first application and second output data from a second application, enabling unified processing or analysis of the combined data. The method may involve normalizing, correlating, or otherwise handling the data to ensure compatibility and coherence between the disparate sources. This approach enhances interoperability between applications, simplifies data workflows, and improves efficiency in systems where multiple applications generate related but separate outputs. The invention is particularly useful in environments where data from different applications must be aggregated, such as in enterprise software, data analytics, or automated workflow systems. By standardizing the input from multiple applications, the method ensures that the data can be processed together, reducing errors and improving decision-making based on consolidated information.
20. The computer-implemented method of claim 13 , further comprising: receiving, from the device, second input data corresponding to a second command; determining second intent data representing the second command; determining third output data based at least in part on the second intent data, the third output data being responsive to the second command; causing the device to present second content corresponding to at least a portion of the third output data; receiving, from the device, third input data corresponding to a third command; determining third intent data representing the third command, the third intent data representing a first frequency, at which additional content is to be output, is to be decreased; and determining a second frequency at which additional content has been output; determining the second frequency is greater than the first frequency; and based at least in part on the second frequency being greater than the first frequency, determining second additional content, nonresponsive to the third command, is not to be output.
This invention relates to a computer-implemented method for processing user commands in a conversational interface system, such as a virtual assistant or chatbot. The problem addressed is managing the output of additional content (e.g., suggestions, recommendations, or follow-up information) in response to user interactions, particularly when the user's intent indicates a preference for reduced additional content. The method involves receiving input data from a device (e.g., a smartphone or smart speaker) corresponding to a command, determining intent data representing the command, and generating output data responsive to the command. The device then presents content based on the output data. Subsequently, the method receives a second command, determines its intent, and generates responsive output data. If a third command is received, the method checks whether the intent includes a request to decrease the frequency of additional content. If the system detects that the current frequency of additional content output exceeds the desired frequency, it suppresses further nonresponsive additional content. This ensures the system adapts to user preferences, reducing unnecessary or excessive suggestions while maintaining responsiveness to direct commands. The method dynamically adjusts content delivery based on inferred user intent, improving user experience by minimizing intrusive or irrelevant outputs.
Unknown
December 29, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.